General comments:

This paper makes a good contribution to the literature on evaluating long-term variability in the wind resource, an ongoing challenge for wind energy science. The authors make good use of a set of long-term (by wind energy standards) tall-tower data sets to compare several methods to estimate multidecadal wind speed variability from much-less-than-multidecadal in-situ observations. They apply several MCP methods and three widely used ML methods to data from sites in different geographical and climatological settings. The ideal outcome would be to determine the best method(s) to use given a particular geographic or climatological setting, but given the complexity of the wind resource, the results are a little more nuanced than that, and understandably so. Nonetheless, the authors describe how the methods compare to each other across these settings and are able to offer some tentative conclusions and recommendations for estimating long-term variability from short-term records.

Specific comments:

(1) Lines 100-105: The ERA5 data set begins in 1940 but your analysis begins in 1950. Given your goal of characterizing long-term variability, why exclude this additional 10 years of reanalysis data?

(2) Table 2: Are all the sites freestanding met towers or are they towers in the vicinity of a wind turbine, which might possibly be affected by wakes from some wind directions?

(3) Line 210: The headings in Table 4 list 2010-2016 and 2012-2016, but the text says the MER is for 1950-2020. I suggest adding that bit of info to the table heading, just as a reminder to readers.

(4) Lines 220-227: I was confused at this point about why you would want to reduce the influence of interannual variability when creating a long-term reference data set. You talk about this a bit later in the paper, but maybe add a note here that you’ll come back to this in section 3.4.1? In general, I think it might be useful to say a little more in the paper about the importance of long-term “data” such as reanalyses, which in theory can include ENSO and other climate patterns that influence long-term variability at a site.

(5) Line 240: You note here, and in several other places in the paper, that the MNER doesn’t seem to depend much on terrain complexity. Do you have any hypotheses as to why? I encourage you to include a bit more discussion of this - even if only possible hypotheses - in the Discussion and Conclusions section (e.g., lines 380-383). Perhaps that discussion can also touch on your comments about sensitivity to sample size and data gaps (lines 259-263).

Technical comments:

(6) Figure 9: The figure caption should say 2012-2016, not 2010-2016.

(7) Figure 10: The figure caption shows 2010-2016 (MNER_2010-2016 %) twice; the second one should be 2012-2016.

Borowski et al. provide an interesting assessment of the impact of inter-annual variability on long-term wind speed estimates. The topic is a valuable one and the authors did a nice job with their research design and analysis. I especially appreciate the helpful tables describing the observational sites and the variety of ML models considered. The manuscript is a bit challenging to follow at times, with references to analyses discussed much later in the text than the reader’s current location, but the findings are nicely summarized in the discussion section.

General comments:

There are a number of extremely short paragraphs (two sentences, sometimes only one) in the text that should either be elaborated upon, combined with other relevant text, or removed if they are found to provide limited context or support for the analysis.

This paper’s value to the wind energy community would benefit significantly if the results were translated into an energy parameter in the discussion. In particular, I could see great value by running the various wind speed time series through a reference turbine power curve and then speaking to the findings in terms of capacity factor ranges.

Specific comments:

Line 29: I recommend removing the word “hindcast”, as this method is often applied forward in time as well, as reference datasets increase their temporal coverage.

Line 40: Diurnal uncertainty is another important component that deserves mention.

Line 50: Concerning the more than 20 methods Lee et al. (2018) compared, can you follow up with a quick mention of which method(s) was found to be most advantageous and why?

Line 90: When aggregating to hourly intervals, are you selecting the top of the hour measurement (which would best align with ERA5’s instantaneous hourly output) or converting to hourly averages?

Line 94: “Modern reanalysis…” – you may wish to update the grammar and intention of this sentence for clarity.

Line 98: This sentence may prove controversial in its current form. Many wind energy researchers feel that a 1-hour temporal resolution is not high and they do not attribute their use of ERA5 to that or the relatively coarse spatial resolution. Rather, many studies show that ERA5 exhibits good correlation with wind observations. You may wish to focus on this aspect, as it ties nicely to MCP.

Line 103: Do you have confirmation that the nearest grid points are on land and not over water, where the dynamics can be very different? Some of your observations are quite close to the coast.

Line 105: Why are you only focusing on 1950 to 2020 when ERA5 has 1940 to present available?

Line 148: This sentence is a bit confusing and implies that temperature and pressure are decomposed into sine and cosine, which Table 3 contradicts.

Section 2.2.2: Again, I encourage use of a different term than “hindcast” as Figure 3 shows that you look forward in time from the measurements as well as behind. Perhaps “Long-term ensemble range”?

Section 3.1: Might align better in Section 2.

Line 205: Can you please be more specific than “see below”? Perhaps guide the reader to a specific section, figure, or table.